exp: cuda ci

.github/workflows/cuda_ci.yml

@@ -0,0 +1,52 @@
+name: cuda-compile
+
+on:
+  push:
+    branches:
+      - main
+  pull_request:
+    branches:
+      - main
+
+jobs:
+  cuda-compile:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v2
+
+      - uses: Jimver/[email protected]
+        id: cuda-toolkit
+        with:
+          cuda: '12.4.1'
+
+      - name: Display CUDA version
+        run: |
+          echo "Installed CUDA version is: ${{ steps.cuda-toolkit.outputs.cuda }}"
+
+      - name: Display CUDA install location
+        run: |
+          echo "CUDA install location: ${{ steps.cuda-toolkit.outputs.CUDA_PATH }}"
+
+      - name: Check NVCC Version
+        run: |
+          nvcc -V
+
+      - name: Install dependencies (Linux no Ceres)
+        run: ./scripts/install_ubuntu_deps_no_ceres.sh
+
+      - name: Install test
+        run: |
+          echo "Install test"
+          mkdir build_dir
+          cd build_dir
+          cmake -DCMAKE_CXX_COMPILER_LAUNCHER=ccache .. -DBUILD_SOPHUS_TESTS=Off -DCMAKE_COMPILE_WARNING_AS_ERROR=On -DSOPHUS_ENABLE_ENSURE_HANDLER=$SOPHUS_ENABLE_ENSURE_HANDLER
+          # Ubuntu builds via Github actions run on 2-core virtual machines
+          make -j2
+          sudo make install
+          cd ..
+          cd examples/cuda
+          mkdir build_dir
+          cd build_dir
+          cmake ..
+          make
+          ls -la

Sophus.code-workspace

@@ -64,7 +64,24 @@
 			"streambuf": "cpp",
 			"thread": "cpp",
 			"cinttypes": "cpp",
-			"typeinfo": "cpp"
+			"typeinfo": "cpp",
+			"bitset": "cpp",
+			"charconv": "cpp",
+			"condition_variable": "cpp",
+			"forward_list": "cpp",
+			"format": "cpp",
+			"mutex": "cpp",
+			"span": "cpp",
+			"variant": "cpp",
+			"__bit_reference": "cpp",
+			"__locale": "cpp",
+			"__threading_support": "cpp",
+			"__verbose_abort": "cpp",
+			"ios": "cpp",
+			"locale": "cpp",
+			"print": "cpp",
+			"queue": "cpp",
+			"stack": "cpp"
 		}
 	}
 }

examples/cuda/CMakeLists.txt

@@ -0,0 +1,25 @@
+cmake_minimum_required(VERSION 3.24)
+
+project(CUDAVectorAdd)
+
+enable_language(CUDA) # Enable CUDA language support
+
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CUDA_STANDARD 14)
+set(CMAKE_CUDA_ARCHITECTURES
+    52
+    60
+    61
+    70
+    75
+    80
+    86)
+
+if(NOT CMAKE_BUILD_TYPE)
+  set(CMAKE_BUILD_TYPE RelWithDebInfo)
+endif()
+
+add_executable(kernel main.cpp kernel.cu)
+
+set_target_properties(kernel PROPERTIES CUDA_SEPARABLE_COMPILATION ON)
+target_link_libraries(kernel PRIVATE cuda)

examples/cuda/kernel.cu

@@ -0,0 +1,32 @@
+#include <cuda_runtime.h>
+#include <iostream>
+
+__global__ void vecAddKernel(float* A, float* B, float* C, int N) {
+    int i = blockDim.x * blockIdx.x + threadIdx than that on x;
+    if (i < N) {
+        C[i] = A[i] + B[i];
+    }
+}
+
+// Wrapper function for the CUDA kernel
+void cudaVecAdd(float* A, float* B, float* C, int N) {
+    float *d_A, *d_B, *d_C;
+    size_t size = N * sizeof(float);
+
+    cudaMalloc(&d_A, size);
+    cudaMalloc(&d_B, size);
+    cudaMalloc(&d_C, size);
+
+    cudaMemcpy(d_A, A, size, cudaMemcpyHostToDevice);
+    cudaMemcpy(d_B, B, size, cudaMemcpyHostToDevice);
+
+    int threadsPerBlock = 256;
+    int blocksPerGrid = (N + threadsPerBlock - 1) / threadsPerBlock;
+    vecAddKernel<<<blocksPerGrid, threadsPerBlock>>>(d_A, d_B, d_C, N);
+
+    cudaMemcpy(C, d_C, size, cudaMemcpyDeviceToHost);
+
+    cudaFree(d_A);
+    cudaFree(d_B);
+    cudaFree(d_C);
+}

examples/cuda/main.cpp

@@ -0,0 +1,33 @@
+#include <cmath>
+#include <iostream>
+#include <vector>
+
+extern void cudaVecAdd(float* A, float* B, float* C, int N);
+
+int main() {
+  int N = 1024;
+  std::vector<float> h_A(N, 0);
+  std::vector<float> h_B(N, 0);
+  std::vector<float> h_C(N, 0);
+
+  // Initialize vectors
+  for (int i = 0; i < N; ++i) {
+    h_A[i] = sin(i) * sin(i);
+    h_B[i] = cos(i) * cos(i);
+  }
+
+  // Call the CUDA kernel wrapper function
+  cudaVecAdd(h_A.data(), h_B.data(), h_C.data(), N);
+
+  // Check the result
+  for (int i = 0; i < N; ++i) {
+    float expected = h_A[i] + h_B[i];
+    if (abs(h_C[i] - expected) > 1e-5) {
+      std::cerr << "Result verification failed at element " << i << "!\n";
+      return EXIT_FAILURE;
+    }
+  }
+
+  std::cout << "Test PASSED\n";
+  return 0;
+}

scripts/install_ubuntu_deps_no_ceres.sh

@@ -0,0 +1,19 @@
+#!/bin/bash
+
+set -x # echo on
+set -e # exit on error
+
+cmake --version
+
+sudo apt update -y
+sudo apt install libc++-dev libgflags-dev libsuitesparse-dev clang
+
+git clone https://gitlab.com/libeigen/eigen.git
+cd eigen
+git checkout c1d637433e3b3f9012b226c2c9125c494b470ae6
+
+mkdir build-eigen
+cd build-eigen
+cmake .. -DEIGEN_DEFAULT_TO_ROW_MAJOR=$ROW_MAJOR_DEFAULT
+sudo make install
+cd ../..